How Datadog Built Bits AI SRE: An Autonomous Incident Investigation Agent That Reduces Time to Resolution by Up to 95%
As distributed systems grow more dynamic and complex, production incidents span more services, involve noisier signals, and generate larger volumes of telemetry data, making it hard for on-call engineers to find root causes quickly.
Early SRE agents performed many tool calls and summarized all telemetry at once, causing token counts to scale linearly with complexity, which degraded model performance and led to incorrect root cause identification when noisy signals distracted the summarization prompt.
Bits AI SRE decreases time to resolution by up to 95% and has received overwhelmingly positive feedback from customers who observed reduced time to root cause detection for complex incidents.
Frequently asked questions
What did this team achieve with this AI workflow?
Bits AI SRE decreases time to resolution by up to 95% and has received overwhelmingly positive feedback from customers who observed reduced time to root cause detection for complex incidents.
What tools did this team use?
Bits AI SRE.
What results were reported?
Time to resolution: up to 95%; Time to root cause detection: reduced time to root cause detection for complex incidents; Agent capabilities improvement: significantly improved over the past year (source-reported, not independently verified).
What failed first in this deployment?
Early SRE agents performed many tool calls and summarized all telemetry at once, causing token counts to scale linearly with complexity, which degraded model performance and led to incorrect root cause identification…
How is this incident management AI workflow structured?
Monitor alert triggers investigation → Hypothesis formulation → Targeted query validation → Recursive sub-hypothesis exploration → Audit-ready root cause analysis.